As I grow from E4 -> E5 and develop people/direction skills, how should I think about and maintain the eng excellence axis?

Keep asking yourself this question after you delivered something with your engineering excellence: "How can I deliver the same impact with half of the time/effort?" You might not reach double productivity any time soon but it's good to always aim high.

Never feel satisfied with "I've done this beautifully". Always seek new approaches to something you already knew very well and prove to yourself that "what an idiot I was for being so inefficient in the past". This may sound unreal to you but it's the path forward.

I've read an engineering blog post talking about having kids. You struggled when you have your first kid. After a while, you get the hang of it and realize how inefficient you were before having any kid. Then the same thing happens again after having a second kid. "Necessity is the mother of invention."

Very often people don't proactively find ways to improve their productivity when it's not necessary, because life is good. Only when they struggle do they have to invent new ways to improve productivity. This E5 promotion looks scary from where you are right now -- holding on to the same level of engineering excellence output and adding extra direction/people output. How is that even possible without burnout? You will have to struggle for a while and then you will get the hang of it.

• Long story short: You need to maintain that meaty E4 code volume as you're moving to E5. In a vacuum, try to stay above 50 diffs/half. Mobile engineers tend to have higher diff counts too.
• The E4 -> E5 promotion is an awkward, transitory period as you need to continue writing a lot of code while also developing E5 skills. Once you get to E5 and stabilize there, you don't need to code as much.
• I have seen stable E5s with only 25 diffs for the half do well. I have never seen an E4 with 25 diffs for the half do well, no matter where they are in their E5 promo journey.
• Have an honest conversation with your manager to set rough guardrails on what a "too low" diff count looks like. Make it clear that you take all numbers with a huge grain of salt, and you deeply understand that you can't reduce SWEs to just numbers.